An SVC-MDC video coding scheme using the multi-core parallel programming paradigm for P2P video streaming

In this paper, we propose a combined SVC-MDC (scalable video coding & multiple description coding) video coding scheme using the multi-core parallel programming paradigm for P2P video streaming, which is denoted Co-SVC-MDC. To date, P2P video streaming applications are widely popular and emphasized, e.g., PPstream and PPlive, because of higher transmission speed and data availability. However, in the heterogeneous P2P network environment, users are able to utilize PDA, notebook or desktop computer through distinct network interfaces to get on-demand videos ubiquitously. To provide distinct spatial-resolution/fidelity videos and flexible video transmission (playback) over P2P networks, a brand new coding architecture needs to be devised. In the proposed Co-SVC-MDC coding scheme, distinct MDC descriptions contain distinct portions of raw video frames, and each raw frame can be compressed as base layer and SVC enhancement layers. In our experiments, a real implementation of Co-SVC-MDC is exhibited and corresponding performances, e.g., PSNR and decoding speeds, are compared with original SVC in distinct congestion-level P2P networks.